AWM: Impact of Intervention Components Table (2021)

Impact of Intervention Components Examined in Sub-Group Analyses

Several sub-group analyses were conducted to investigate the relative effect of interventions with different delivery methods compared to controls. While adjusting results for multiple sub-group analyses is not suggested for systematic reviews, authors recognize that conducting many analyses precludes firm conclusions from any single result. In addition, authors recognize that comparing effect sizes does not imply that the differences are statistically or clinically significant. However, examining patterns in data according to effect sizes compared to controls may explain some heterogeneity in results and inform practitioners. (see Table below).

In-Person or Telehealth vs Controls

Interventions delivered exclusively in-person resulted in a significant improvement in all five outcomes for which sub-group analyses were conducted compared to controls. Interventions delivered exclusively remotely had a smaller effect size compared to controls than their counterparts for the outcomes of BMI and WC and did not significantly improve SBP or FBG compared to controls. There was overlap in confidence intervals. Results from hybrid interventions were more like those delivered in-person (Table).

Individual or Group Intervention vs Controls

Interventions delivered in both the individual and group settings combined resulted in significant improvement in all outcomes compared to controls in which sub-group analyses were conducted. Neither interventions delivered exclusively in a group or exclusively individually resulted in significantly improved FBG levels compared to controls and interventions delivered exclusively individually did not significantly improve SBP compared to controls (Table).

Dietitian Alone or Multidisciplinary Intervention vs Controls

All interventions included a dietitian, but some interventions additionally included other health professionals. Effect size was similar between intervention types compared to controls for anthropometric outcomes, but interventions focusing on nutrition alone had a greater improvement on SBP and FBG compared to controls, though confidence intervals overlap (Table).

Number of Contacts with a Dietitian

Interventions including at least five contacts with the dietitian had a larger effect size, compared to controls, than interventions with four or less contacts for anthropometric measures, SBP and FBG. However, there was overlap in confidence intervals. Interventions with four or less contacts did not result in significantly improved SBP or FBG.

Frequency of Contacts with the Dietitian

There were no clear dose-response patterns between frequency of contacts with the dietitian and outcomes, except for FBG, for which greater frequency resulted in a greater effect size, compared to controls, then lower frequency interventions, though confidence intervals overlapped.

Study Duration

For nearly every outcome examined, studies that were at least 12 months in duration resulted in the largest effect size compared to controls (effect sizes were similar with studies of 6-12 months duration for percent weight loss), though there was overlap in confidence intervals. Studies less than 12 months did not result in significantly improved SBP or FBG compared to controls.

Table. Results of Sub-group Analysis According to Intervention Characteristics [Mean Difference (95% CI)}.^a(Download Table)

	BMI (kg/m²)	% Weight Loss	Waist Circumference	Systolic Blood Pressure (mmHg)	Fasting Blood Glucose^b (mg/dL)
In-Person or Telehealth Interventions vs Controls
Exclusively In-Person	-1.71 (-2.05, 1-.38)	-3.59 (-5.20, -1.99)	-3.60 (-4.57, -2.64)	-4.67 (-6.89, -2.45)	-1.20 (2.38, -0.02)
Exclusively Remote	-0.82 (-1.34, -0.30)	-5.17 (-7.73, -2.61)	-1.94 (-3.72, -0.15)	1.59 (-1.55, 4.73)	-0.79 -5.29, 3.72)
Blended In-Person and Remote	-1.16 (-1.45, -0.87)	-4.56 (-6.24, -2.89)	-3.60 (-5.22, -1.98)	-1.67 (-3.81, 0.46)	-2.74 (-4.85, -0.63)
Individual or Group Interventions vs Controls
Exclusively Individual	-1.28 (-1.56, -1.00)	-2.96 (-3.99, -1.94)	-4.06 (-5.61, -2.52)	-0.76 (-3.20, 1.67)	-0.02 (-2.12, 2.07)
Exclusively Group	-2.15 (-3.41, -0.89)	-4.88 (-7.95, -1.81)	-3.00 (-4.41, -1.59)	-8.10 (-12.96, -3.25)	-4.46 (-9.78, 0.87)
Blended	-1.07 (-1.36, -0.79)	-4.19 (-6.49, -1.90)	-2.67 (-3.61, -1.74)	-2.01 (-3.53, -0.50)	-3.24 (-5.59, -0.89)
Dietitian Alone or Multidisciplinary Interventions vs Controls
Dietitian Alone	-1.49 (-1.86, -1.13)	-3.81 (-5.03, -2.60)	-3.03 (-4.3, -1.88)	-4.57 (-7.17, -1.98)	-3.20 (-6.19, -0.21)
Multidisciplinary	-1.48 (-1.82, -1.15)	-4.93 (-7.10, -2.77)	-3.70 (-4.88, -2.53)	-0.84 (-3.50, 1.82)	-1.84 (-3.94, 0.27)
Number of Contacts with Dietitians vs. Controls
1-4 Contacts	-1.46 (-1.84, -1.08)	-2.36 (-3.08, -1.63)	-2.46 (-3.29, -1.64)	-0.95 (-3.76, 1.86)	0.31 (-7.56, 8.18)
≥ 5 Contacts	-1.55 (-1.88, -1.22)	-4.23 (5.71, -2.74)	-3.63 (-4.67, -2.59)	-3.34 (-5.58, -1.10)	-1.44 (-2.85, -0.01)
Frequency of Contacts with Dietitians vs Controls
<1/month	-1.96 (-2.93, -0.99)	-3.39 (-4.79, -1.98)	-2.75 (-5.08, -0.42)	-2.45 (-5.59, 0.69)	1.80 (-4.81, 8.41)
1-3/month	-1.38 (-1.66, -1.10)	-4.26 (-6.25, -2.26)	-3.85 (-4.97, -2.74)	-3.53 (-6.45, -0.62)	-1.16 (-2.61, 0.29)
≥4/month	-1.60 (-2.26, -0.95)	-3.91 (-5.75, -2.06)	-2.02 (-3.23, -0.80)	-2.01 (-4.80, 0.79	-4.24 (-7.92, -0.57)
Intervention Study Durations vs Controls
<6 months	-1.50 (-1.89, -1.10)	-3.04 (-4.13, -1.94)	-3.21 (-4.43, -1.98)	-0.80 (-2.91, 1.32)	-3.15 (-6.39, 0.09)
6<12 months	-1.11 (-1.56, -0.67)	-4.57 (-8.06, -1.09)	-2.81 (-4.71, -0.92)	-4.11 (-8.56, 0.33)	-0.59 (-3.61, 2.42)
≥12 months	-1.94 (-2.48, -1.40)	-4.36 (-6.24, -2.49)	-4.12 (-5.18, -3.07	-3.03 (-4.60, -1.46	-1.77 (-3.22, -0.32)

^aAll results are compared to control groups. The outcomes of quality of life and cost-effectiveness were reported in too few studies for results to be stratified in sub-group analyses.

^bFor participants with no Type 2 Diabetes.

updated: 07/29/2021